As others has stated dot gain can be effected by various processes along the production path. The problems we had were initially blamed on the press operator and he was told to reprint, but despite rechecking everything at his end the tint was still too dark. We eventually tracked it down to the CTP machine, it was over exposing the plate. Apparently as the components age, the recommended calibration every 6 months we were having was not enough to keep it correctly calibrated. Other issues that gave us grief was old ink being used to print the job originally and new ink being used to print a reorder, a bad batch of plate material, customers colour management profiles causing 'odd' colour shifts due to conversion, customers choice of tint values, with our presses anything over 70% starts to look like 95% and with blue pantone inks we generally reduce the tint by 3-5% if we want to allow for the increase in dot gain. One of the ways we tried to combat the problem was to add a strip of light to dark labelled tints (CMYK or Pantone Spot) at the bottom of the sheets with the stock info on. The finisher would trim these off and known good samples were kept and used for comparison against future print jobs. Or when customers changed from X coated to Y uncoated stock the tints could then be compared and adjustments made to meet customers expectations. The rule of thumb I use is to presume it's prepress first. If your press operators have been with you for a couple of years, the chances are they will know when something is up at their end and say so. While a lot of prepress problems can be effectively invisible until its actually on the press. Hope you manage to track it down.
